Teardown of Kathrein 1800/900 MHz Antenna For Mobile Phone Base Station

[Mads Barnkob]

Teardown of a Kathrein Dual Panel 1800/900 MHz used for both GSM (2G) and 3G mobile broadband technologies. It has options for Remote Electrical Titlt (RET) and on the inside we will find phase changers, splitters, combiners and 1800/900 MHz antenna elements. I give a walk-through of the circuit, what the different elements does and what kind of antenna design it is based upon.

/>
The input signal goes through a signal splitter where the 1800 and 900 MHz signals are splitted or combined. Pure black magic, I can not tell you what is going on here... 



The base / back plate of the antenna is based on the slanted sidewall principle.



The antenna array consists of both 1800 and 900 MHz elements, more are used where some are static and some are fed through the remote electrical tilt option, in order to be able to change the shape of the antenna lobe.



The Remote Electrical Tilt function (RET) is a simple motor adjusting a isolated rod to the phase shifter mechanism, that can change phasing of the signal by 15 degrees.



all coaxial wiring is made from silvered wire and kept in either red or white insulating material to distinguish between the 1800 and 900 MHz signals.

[Da_Stier]

Very interstesting look at this antenna.

To start my few comments on this thing, I want to start by copying my comment from the youtube video:

"Very interesting, I love how every panel like this (every part number) seems to be built a little bit different with different elements and a different architecture.
I have some of the bigger dipole elements with the more common "small sqares" in the middle.

Also note how every color of semi rigid coax is a different impedance.
Kathrein loves to use coax as transmission line transformers for matching all the different components.

The coaxes are looped around the inside of the unit to keep the phase matched up, since every feedline needs to be the same - or atleast some known - length.

The input diplexers are a piece of art and very stable.
I wonder if anyone could come up with a design, that is as simple and effecitve.
I guess these stripline designs are all autogenerated or atleast auto matched for efficiency.

As Robberiffic said, they are a combination of filters and a power splitter.
The upper dielectric makes them embedded striplines and also gives a higher dieelectric constant inside the enclosure.
With a big airspace it is quite hard to keep oscillations down in such a enclosure with such high powers.
All in all this sector antenna seems to be built quite a bit more solid than the ones I have seen. (also from Kathrein)"

This is the element I was talking about:
Unfortunately I can't remember what part number the sector antenna had and my phone with all the teardown pictures for the antennas died last year.



If you are interested, I can take some S parameter measurements of the Diplexer unit and or the elements, if you still have them around.




Greetings,
Michael

[Da_Stier]

So I have some more intersting bits and pieces to share.

On a Kathrein panel that was pretty similar to yours, the input diplexer was made from 3 parts.
The power splitter and a filter for each band.

The splitter was pretty basic and not very interesting.
The filters on the other hand were similar to yours.
A stripline filter realised as an embedded stripline inbetween two substrates.

This is the filter for the lower band.
Interestingly it is the smaller one of the two.
However it is more of a lowpass filter and not a bandpass.



I don't have my GPIB network setup at the moment so I could only take a literal screenshot of the VNA.



This is the filter for the highband.
It can be used in the 1800MHz band, as well as the 2100MHz band.



With this one you can actually see the filter structure, made from stripline stubs, if you look at it very closely.
Excuse my paint skills but I guess it shows what I meant.



And the S21 response of the filter:



These filters have a surprisingly sharp cutoff and pretty good isolation outside its passbands.
I would love to see, how these filters are made by Kathrein, my bet is on pretty much automated simulation to optimize them.

EDIT: At around 11:40 in your video you showed a lot of figures showing sector antenna elements and their responses. What is that from, I would be interested in taking a look at that paper.



Greetings,
Michael

[Mads Barnkob]

Thanks for all the new information, I got a few things wrong in the video, so its great there is always something to learn

I will send you an email about the diplexer and antenna elements, so you can have some fun with those.

The paper on antenna simulations is here: https://www.researchgate.net/publication/327153500_A_Miniaturized_Base_Station_Antenna_with_Novel_Phase_Shifter_for_3GLTE_Applications looks like it should have been for IEEE members only, but its publicly available on researchgate.. oh well

[Da_Stier]

Thanks for all the new information, I got a few things wrong in the video, so its great there is always something to learn

I think you did a very good job on the video.
Especially for people who might have never seen one of these antennas, it is a great amount of information.
Your first antenna video was btw the first time I saw the inside of one of these sector antennas and it blew my mind back then.

I will send you an email about the diplexer and antenna elements, so you can have some fun with those.

Sounds great 

The paper on antenna simulations is here: https://www.researchgate.net/publication/327153500_A_Miniaturized_Base_Station_Antenna_with_Novel_Phase_Shifter_for_3GLTE_Applications looks like it should have been for IEEE members only, but its publicly available on researchgate.. oh well

Thanks, looks very interesting after a first look, I guess I will have something to read.


One thing that I forgot to mention previously, your panel seems to be made quite a bit more solidly.
The antenna I took apart was made from thin tin plated copper sheetmetal instead of the thick aluminium profile.
Also the filters are made from two PCBs that are glued and or soldered together, instead of the nice diplexer unit from your antenna.



Greetings,
Michael

[Da_Stier]

Hi all,

today I took a closer look at the diplexer units.
The first thing I noticed was, how heavy they are, they are actually made of solid milled brass.
So this antenna seems to be either an earlier one or a higher quality one, since some other Kathrein antennas I have seen use Diplexers made from PCB only, without a metal case.

The next fascinating thing was that the "traces" in the diplexer aren't etched onto the lower substrate but are a very thin lasercut metal sheet instead.
By unsoldering the coax connections, you can actually take it out seperately.

The substrate seems to be some kind of teflon based RF substrate. It has this kind of destinctive "wet" feel of teflon to it.
Maybe slick is the better word for it?

So I decided to connectorize one of them to make it a nice reusable RF block.
To do so I first disassembled the unit.
At first I wanted to get the connectors onto the side, however the unit is slightly to thin to be able to mount them in a good way.
So instead I decided to mount them on the bottom.
So I drilled out some holes in the bottom brass piece and the bottom substrate layer.
I also milled some small pockets to help aligning it and also to make sure that the center pins of my SMA connectors are long enough.
This is the unit after modification:



Next I soldered on the SMA connectors and continued by assembling the unit again.
This is the finished diplexer with connectos:



Like this it is really easy to use it for experiments but also to connect it up to the VNA.
This is the S21 response of the lower band, which is basically from 760MHz to around 1GHz and therefore covers all the lower cellular bands:



This is the S21 response of the high band, which is basically from 1.6GHz to 2.3GHz wich covers all the high cellular bands (except the 2.6GHz band, which I have never seen around Europe very much, even though it should be used here)



The responses look very clear and defined with very little insertion loss.
I might connectorize the second one as well.





I hope you liked this look at the diplexers,
Michael